Answer
Should you tell clients you use AI?
Where it touches their information or changes what they receive, yes. Where it is a tool you used, it is a tool.
Where their information went elsewhere, or where what they receive differs materially from what they think they are buying, yes. Where it is a tool you used to do the work you were engaged for, it is a tool.
The instinct is to treat this as a question about honesty in general, which produces either blanket disclosure that means nothing or silence that feels uncomfortable. The productive framing is narrower: what would a reasonable client want to know because it affects them, rather than what is technically true about how the work was produced.
That yields two triggers. The first is their information. If material a client gave you went to an external service, that is a disclosure regardless of how routine the task was, because they made a decision about who holds their information when they engaged you. For anything under a confidentiality obligation this is not a courtesy question at all — the agreement decides it, and the agreement may prohibit it.
The second is the nature of what they receive. A client engaging you for judgement, review or original analysis is buying your attention, and a deliverable produced substantially without it is not the thing they bought whether or not it is good. Where the tool changed what the work is rather than how fast it was done, that is the case for telling them, and it is also usually the case for reconsidering the arrangement.
Where neither applies, the tooling is tooling. Drafting from your own notes, restructuring your own analysis, checking your own work, producing a first version you then rewrite: these are how the work got done and clients do not generally want an inventory of methods. Disclosing them is not more honest, it is less informative, because it buries the two things that would have mattered under a list of things that do not.
Regulated professions are a separate matter and the rules may be explicit. Some jurisdictions require professionals to disclose automated interaction proactively when providing regulated services, and several professional bodies have addressed client consent for processing client material. Where you hold a licence, the question has an answer supplied by somebody else, and it should be established rather than reasoned from first principles.
There is a commercial argument worth separating from the ethical one. Some clients will react badly and some will react well, and both reactions are usually about what they think it implies rather than about the tool. A firm that can explain what it uses these tools for, what it does not use them for, and what checking happens tends to have a better conversation than one that either avoids the subject or announces it without those specifics.
Nobody discloses their spellchecker; the question is whether this is a spellchecker or a subcontractor.
Siddharth Sharma, Context Theory
Related questions
What if a client asks directly?
Answer accurately, and answer with the specifics: what it is used for, what it is not used for, and what review the output receives. A yes or a no is a worse answer than either, because the question is really about how the work is done and both bare answers leave that unaddressed.
Should it be in the engagement terms?
Where client material may be processed externally, yes, because that is a permission rather than a notification and getting it in advance is much easier than getting it afterwards. Once it is in the terms, the ongoing disclosure question largely disappears, which is the practical reason to do it.
METHOD
Every figure below carries its source and the date it was verified. Nothing on this page is asserted.
The numbers on this page.
| What | Value | Specific to |
|---|---|---|
| US SMB retainer, focused one-to-two-service engagement | $1,500–$4,000 | Category-wide |
| Share of the buying journey completed before contacting a vendor | 60% | Category-wide |
2026 agency pricing survey · per month · verified
2026 B2B buyer surveys · verified
What is specific to this page.
| Kind | Claim | Check it against |
|---|---|---|
| Constraint | Where client material is processed by an external service, the disclosure question is decided by the engagement and any confidentiality obligation rather than by the routineness of the task, because the client made a decision about who holds their information. | The confidentiality and subcontracting provisions in the engagement terms. |
| Procurement | A client engaging for judgement, review or original analysis is purchasing attention, so a deliverable produced substantially without it differs from what was bought regardless of quality, which is the second disclosure trigger. | Comparing what the engagement describes as the deliverable against how it was actually produced. |
| Licensing | Some jurisdictions require licensed professionals to disclose automated interaction proactively when providing regulated services, so for licence holders the question is settled externally rather than by judgement. | The disclosure provisions applicable to the profession in the jurisdictions where services are provided. |
| Buying behaviour | Blanket disclosure of every method buries the two consequential facts among inconsequential ones, which makes it less informative to the client rather than more honest. | Comparing what a client understands from a general statement of tool use against a specific one covering their material and their deliverable. |
Each row would be wrong on another industry's page. Where a sourced figure exists it is in the table above instead; these are the constraints that shape the work and do not happen to be numbers.
Start with the measurement.
Reading about a benchmark is not the same as knowing your own number. The audit produces yours, measured rather than estimated.
$497 · delivered in 5 business days · credited against month one